CHEESE Modeller

ADMET Prediction | QSAR custom model training | Acceleration of Docking

Ultra Fast Prediction
Everyone can use it
Your Data & Your Models

  • Ultra-Fast ADMET Prediction

  • Train Custom QSAR in 2 Minutes

  • The Simplest GUI on the Market

  • Accessible to Non-Experts

  • Automatic Parameter Selection

  • Single-Click Training

  • Interpretable Evaluation and Validation

  • Data Storage for your Molecules

  • 2000+ Curated Activity Assays for QSAR

  • 25+ Ready-to-use ADMET Models

Virtual Screening Case Study
ADMET Benchmark
Clustered Prediction

  • 147x Improved Hit Rate at 10nM from 0.02% to 2.93% over random selection

  • Case study from our customer: Bridgene Biosciences

  • Achieved same success rate in 1 day as the customer would get with 200 days of docking

  • Competitive Results in 22 ADMET Benchmarks (Therapeutic Data Commons)

  • Outperforms SOTA in 4 cases

  • Just a simple linear model on top of CHEESE Embeddings (Electrostatic)

  • Without hyperparameter tuning

  • Clustered Prediction allows fast property filtering of large databases

  • 4-7 orders of magnitude faster than traditional prediction

  • Keeps 80% of the accuracy

Predict ADMET Properties
Train QSAR Models
Accelerate Docking

On-Prem

100% Privacy | Behind Your Firewall | Integrate with proprietary data

  • check_circle

    CHEESE Modeller On-Prem can be installed privately in your company

  • check_circle

    Seamless integration with your datasets and with other tools in CHEESE Platform

  • check_circle

    Company-wide license: Fixed price, No limits on the number of users

Services

QSAR Model Development

Do you already have some measured activities (is its just a dozen, no problem)? Using the CHEESE Modeller platform, we can develop quantitative structure-activity relationship models that reveal patterns between molecular structures and their biological activity.

ADMET Property Prediction

Our experts use the CHEESE Modeller platform to train models that accurately predict ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) properties. This critical step helps you assess the safety and efficacy of your drug candidates, streamlining the path from discovery to development.

Lead Optimization & Active Learning

Our lead optimization services leverage QSAR and ADMET models to predict and refine the properties of newly designed molecular modifications. We are able as well to activelly update the AI models based on docking or experimental results to maximize the VS performance.

Collaboration with Deep MedChem has significantly streamlined the implementation of AI/ML technologies for processing chemoinformatics data in our drug discovery processes. Through a series of open consultations, the members of Deep MedChem quickly and effectively understood our needs. As a result of this collaboration, we now possess several proprietary models that provide us with significant competitive advantage over other solutions.
Jan Skácel
Research Group Lead, IOCB Prague
CHEESE Modeller has been a game-changer for our drug discovery projects.
The tool accurately predicted docking scores in our virtual screening saving us 200 days of computation time, while achieving unprecedented hit rate. Its simplicity and ease of use set it apart from competitors. Additionally, their customer support is outstanding, always providing prompt and effective assistance including updates in their product. Collaborating with Deep MedChem has boosted our research capabilities and competitive advantage.
Chief Scientific Officer
A San Francisco Bay Area Biotech Company

Contact us

We look forward to helping you make your project a success!